lkml.org 
[lkml]   [2018]   [May]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 3/3] sched/fair: add tracepoints for cfs throttle
Date
This patch introduces two sched tracepoints to track
CFS throttle and unthrottle. The use case is to measure
how long each throttle is and to track when a CPU cgroup
is throttled and unthrottled.

Sample output:

cfs-722 [000] dN.3 51.477702: sched_cfs_throttle: path=/test cpu=0 runtime_remaining=0
<idle>-0 [000] d.h4 51.536659: sched_cfs_unthrottle: path=/test cpu=0 runtime_remaining=1

Cc: Paul Turner <pjt@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
include/trace/events/sched.h | 42 ++++++++++++++++++++++++++++++++++++++++++
kernel/sched/fair.c | 2 ++
2 files changed, 44 insertions(+)

diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index bc01e06bc716..3d9e00972e92 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -570,6 +570,48 @@ TRACE_EVENT(sched_wake_idle_without_ipi,

TP_printk("cpu=%d", __entry->cpu)
);
+
+#ifdef CONFIG_CFS_BANDWIDTH
+DECLARE_EVENT_CLASS(sched_fair,
+
+ TP_PROTO(struct cfs_rq *cfs_rq),
+
+ TP_ARGS(cfs_rq),
+
+ TP_STRUCT__entry(
+ __field( s64, runtime_remaining )
+ __field( int, cpu )
+ __dynamic_array(char, cfs_path,
+ cgroup_path(cfs_rq->tg->css.cgroup, NULL, 0) + 1)
+ ),
+
+ TP_fast_assign(
+ __entry->runtime_remaining = cfs_rq->runtime_remaining;
+ __entry->cpu = cpu_of(cfs_rq->rq);
+ cgroup_path(cfs_rq->tg->css.cgroup,
+ __get_dynamic_array(cfs_path),
+ __get_dynamic_array_len(cfs_path));
+ ),
+
+ TP_printk("path=%s cpu=%d runtime_remaining=%lld", __get_str(cfs_path),
+ __entry->cpu, __entry->runtime_remaining)
+);
+
+DEFINE_EVENT(sched_fair, sched_cfs_throttle,
+
+ TP_PROTO(struct cfs_rq *cfs_rq),
+
+ TP_ARGS(cfs_rq)
+);
+
+DEFINE_EVENT(sched_fair, sched_cfs_unthrottle,
+
+ TP_PROTO(struct cfs_rq *cfs_rq),
+
+ TP_ARGS(cfs_rq)
+);
+#endif /* CONFIG_CFS_BANDWIDTH */
+
#endif /* _TRACE_SCHED_H */

/* This part must be outside protection */
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 8a13ee006f39..3bcc40f2f272 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4736,6 +4736,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)

cfs_rq->throttled = 1;
cfs_rq->throttled_clock = rq_clock(rq);
+ trace_sched_cfs_throttle(cfs_rq);
raw_spin_lock(&cfs_b->lock);
empty = list_empty(&cfs_b->throttled_cfs_rq);

@@ -4766,6 +4767,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
se = cfs_rq->tg->se[cpu_of(rq)];

cfs_rq->throttled = 0;
+ trace_sched_cfs_unthrottle(cfs_rq);

update_rq_clock(rq);

--
2.13.0
\
 
 \ /
  Last update: 2018-05-22 08:21    [W:0.066 / U:1.672 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site